Brookfield and NextEra Plan $100 Billion AI Campus at Former DOE Site With Dedicated Power
A massive 1.2-gigawatt data center project in Kentucky aims to solve the AI power crisis by building its own dedicated natural gas and battery storage infrastructure.
How this story has developed
This report is part of a developing story — read the earlier chapters below.
- Global Data Center Power Consumption Reaches 2% of World Electricity and 6% of US Grid
- Is the AI Boom a Hidden Tax on Your Utility Bill and the Climate?
- Brookfield and NextEra Plan $100 Billion AI Campus at Former DOE Site With Dedicated Power (this article)
- Infrastructure Developers
- Argue that bringing dedicated power is the only sustainable way to scale AI without burdening communities.
- Ratepayer Advocates
- Focus on the promise that residential customers won't foot the bill, emphasizing the Ratepayer Protection Pledge.
- Energy Analysts
- Point out the technical challenges of matching battery storage with continuous AI compute demands, questioning the nameplate capacity numbers.
How we got here
1952
The Paducah Gaseous Diffusion Plant is established to enrich uranium for national defense during the Cold War.
2013
Commercial uranium enrichment operations cease, and the site transitions to long-term environmental remediation.
Nov 2025
The Department of Energy issues a Request for Offers to redevelop portions of the site into an American Energy Hub.
Jul 2026
Brookfield and NextEra Energy are selected to build a $100 billion AI data center and power campus.
2028
Targeted date for the initial phase of data center operations to begin.
2032
Expected completion of the full 1.2-gigawatt compute campus and 4.6-gigawatt power infrastructure.
Why it matters
As artificial intelligence models demand unprecedented amounts of electricity, local communities are increasingly bearing the cost of grid upgrades. This project pioneers a 'bring-your-own-power' model that could allow the U.S. to scale its AI infrastructure without driving up residential utility bills.
The artificial intelligence boom is threatening to drive up your electricity bill. As massive data centers demand unprecedented amounts of power to train and run frontier models, local utilities have struggled to keep pace. In many regions, the cost of upgrading the electrical grid to support these hyperscale facilities threatens to fall on residential ratepayers, sparking political pushback and fears of regional energy shortages.[1][4]
Now, a $100 billion project in western Kentucky is attempting to rewrite the blueprint for AI infrastructure. Brookfield and NextEra Energy have partnered to transform a former Cold War uranium enrichment site into a massive artificial intelligence campus that brings its own electricity. Backed by the U.S. Department of Energy, the Paducah American Energy Hub represents one of the largest private investments in digital infrastructure in American history.[1][2]
The scale of the proposed campus dwarfs traditional data centers. When fully constructed, the site will support more than 1.2 gigawatts of compute capacity and up to 1.8 gigawatts of total utility capacity. To put that in perspective, a single gigawatt is roughly equivalent to the output of a traditional nuclear reactor, capable of powering about 750,000 homes.[1][6]
The core innovation of the Paducah project is its "bring-your-own-power" architecture. Instead of simply plugging into the existing grid and draining local resources, the data center will be powered by up to 4.6 gigawatts of dedicated generation and storage built specifically for the site. This model shifts the burden of power generation from public utilities to private developers.[4][5]

To provide the continuous, uninterrupted "baseload" power that AI workloads demand, NextEra Energy plans to construct 2 gigawatts of natural gas-fired generation. Because AI training runs cannot be paused when the wind stops blowing or the sun goes down, natural gas provides a reliable, dispatchable energy source that ensures the massive clusters of specialized microchips remain operational around the clock.[1][6]
Alongside the natural gas plants, the developers will deploy up to 2.6 gigawatts of battery energy storage systems. These massive battery banks act as a buffer, storing excess electricity during periods of low demand and discharging it during peak hours. This hybrid approach allows the campus to manage its massive energy appetite without causing sudden spikes or voltage drops on the surrounding electrical grid.[1][3]
Because the dedicated power generation exceeds the 1.2 gigawatts of compute capacity required by the data center, the campus will not strain the local grid. In fact, the surplus electricity generated by the natural gas and battery systems will flow back into the regional network, effectively turning the data center into a net contributor to the area's power supply.[2][6]
Because the dedicated power generation exceeds the 1.2 gigawatts of compute capacity required by the data center, the campus will not strain the local grid.
This self-contained design fulfills the Trump administration's Ratepayer Protection Pledge. By ensuring that the multibillion-dollar infrastructure costs are borne entirely by Brookfield and NextEra, the project insulates existing residential and small-business customers from rate hikes. Energy Secretary Chris Wright highlighted this as a crucial roadmap for building world-class infrastructure without passing costs onto surrounding communities.[1][4]
The choice of location is highly deliberate and solves a major bottleneck in data center construction. The Department of Energy's Paducah Gaseous Diffusion Plant enriched uranium for national defense from 1952 until commercial operations ceased in 2013. While the 3,500-acre site requires ongoing environmental remediation, it already possesses the critical arteries of heavy industry.[3][5]

Repurposing this federal "brownfield" land shaves years off the typical development timeline. The site is already equipped with high-voltage transmission lines, massive water cooling infrastructure, and extensive fiber-optic connectivity. In an era where acquiring land and securing grid interconnection can delay projects by half a decade, the Paducah site offers a ready-made foundation for hyperscale computing.[3][6]
To integrate this massive influx of power and data, the developers have formed a coalition with three regional utilities. Big Rivers Electric Power Corporation will manage wholesale electricity service, Jackson Purchase Energy Cooperative will handle retail delivery, and Paducah Power System will act as a community partner. The power service agreement remains subject to final approval from the Kentucky Public Service Commission.[5][6]
Beyond the technical milestones, the project promises a significant economic revitalization for western Kentucky. The coalition estimates the development will create approximately 8,000 construction jobs and 600 permanent operations positions. Local officials view the campus as a generational opportunity to replace the jobs lost when the uranium enrichment facility shuttered over a decade ago.[1][3]
Despite the massive scale, energy analysts caution that nameplate capacity numbers require careful scrutiny. A gigawatt of battery storage cannot provide the continuous power that AI workloads demand in the same way a natural gas plant can. Experts note that the project will rely heavily on its fossil-fuel components to maintain uptime, emphasizing that battery capacity is not synonymous with continuously available generation.[4]
This reliance on natural gas also raises questions about the long-term environmental footprint of the campus. While the project solves the immediate grid capacity problem, the emissions profile of a 2-gigawatt gas plant runs counter to the zero-carbon pledges made by many major technology companies. The coalition has not yet announced which anchor tenant will lease the compute capacity, leaving it unclear how the power mix will align with corporate climate goals.[4][5]
Construction is slated to begin following regulatory approvals, with initial data center operations targeted for 2028 and full buildout expected by 2032. As the AI industry continues to collide with the physical limits of the American electrical grid, the Paducah campus will serve as a high-stakes test case. If successful, the model of pairing hyperscale compute with dedicated, privately funded power generation could become the standard for the next era of digital infrastructure.[2][5]
What to know
- Brookfield and NextEra Energy plan to build a $100 billion AI data center campus at a former Department of Energy site in Kentucky.
- The facility will feature a 'bring-your-own-power' model, utilizing up to 4.6 gigawatts of dedicated natural gas and battery storage.
- The project is designed to protect local ratepayers by ensuring infrastructure costs are not passed onto residential utility bills.
- Surplus electricity generated by the campus will flow back into the regional grid.
- The development is expected to create 8,000 construction jobs and 600 permanent operations roles by its 2032 completion.
Where opinion splits
The Infrastructure Developers' View
Proponents argue that massive, self-powered campuses are the only viable path forward for American AI dominance.
Brookfield and NextEra view the Paducah project as a blueprint for the future of digital infrastructure. By pairing hyperscale compute directly with dedicated natural gas and battery storage, developers can bypass the years-long interconnection queues that currently plague the public grid. They argue this "bring-your-own-power" model allows the U.S. to maintain its lead in artificial intelligence without triggering regional energy crises or forcing local utilities to ration electricity.
The Ratepayer Advocates' View
Consumer groups and policymakers focus on insulating local residents from the multibillion-dollar costs of grid upgrades.
For years, local communities have pushed back against data centers that drain regional power supplies and drive up residential utility bills. Advocates view the Paducah agreement as a major victory, pointing to the Trump administration's Ratepayer Protection Pledge as a structural safeguard. Because the campus will generate more power than it consumes and flow the surplus back to the grid, proponents argue it transforms a potential community burden into a net benefit for western Kentucky.
The Energy Analysts' View
Skeptics caution that the massive nameplate capacity numbers obscure the technical realities of powering AI.
While the headline figures of 4.6 gigawatts of generation are impressive, energy analysts warn against equating battery storage with continuous baseload power. AI training clusters require uninterrupted electricity, meaning the campus will rely heavily on its 2 gigawatts of natural gas generation to maintain uptime. Analysts question whether the project can truly balance its massive fossil-fuel footprint with the aggressive zero-carbon mandates of the tech giants likely to lease the space.
Key terms
- Hyperscale Data Center
- A massive computing facility designed to support the enormous processing requirements of cloud computing and artificial intelligence.
- Gigawatt (GW)
- A unit of power equal to one billion watts, roughly equivalent to the output of a large nuclear reactor or enough to power 750,000 homes.
- Bring-Your-Own-Power (BYOP)
- An infrastructure model where large industrial facilities build their own dedicated electricity generation rather than relying solely on the public grid.
- Battery Energy Storage System (BESS)
- Large-scale battery installations that store excess electricity and discharge it when demand is high or generation is low.
- Ratepayer
- A residential or commercial customer who pays for utility services, often vulnerable to price increases when grid upgrades are required.
Unanswered questions
- The identity of the anchor tenant: The coalition has not announced which major tech company or AI lab will lease the 1.2 gigawatts of compute capacity.
- The environmental impact: The exact emissions profile of the 2-gigawatt natural gas generation component remains unclear, as does the timeline for potential environmental reviews.
- Battery feasibility at scale: Analysts question whether 2.6 gigawatts of battery storage can economically provide the continuous, uninterrupted power required by AI workloads.
Reader questions
Will this project increase local electricity bills?
No. The developers have committed to a 'bring-your-own-power' model, meaning the data center will pay for its own dedicated natural gas and battery infrastructure rather than passing costs to existing ratepayers.
Why build this on a former uranium enrichment site?
The Department of Energy's Paducah site already possesses extensive high-voltage transmission lines, water infrastructure, and fiber connectivity, which significantly reduces the time needed to build a new campus.
How many jobs will the project create?
The coalition estimates the development will generate approximately 8,000 construction jobs and 600 permanent operations positions once fully built.
How much power will the campus actually use?
The data center is designed to support 1.2 gigawatts of compute capacity, but the developers are building up to 4.6 gigawatts of dedicated generation and storage to ensure reliability and supply surplus power to the grid.
Sources
[1]BloombergRatepayer Advocates
NextEra, Brookfield Plan $100 Billion AI Campus at DOE Site
Read on Bloomberg →[2]MorningstarRatepayer Advocates
Brookfield, NextEra Team to Build $100 Billion AI Data Center
Read on Morningstar →[3]Engineering News-RecordEnergy Analysts
DOE Taps Brookfield, NextEra for $100B AI Campus in Kentucky
Read on Engineering News-Record →[4]AI WeeklyEnergy Analysts
Brookfield Wins DOE Paducah Site for $100B AI Campus Plan
Read on AI Weekly →[5]Converge DigestInfrastructure Developers
Brookfield, NextEra to Develop 1.2 GW AI Campus at Paducah Site
Read on Converge Digest →[6]The Data Center EngineerInfrastructure Developers
DOE selects Brookfield, NextEra for 1.8 GW AI data center campus in Kentucky
Read on The Data Center Engineer →
Comments
Every angle. Every day.
Get ai stories with full source coverage and perspective breakdowns delivered to your inbox.








